Leaf senescence is the final stage of leaf development and is finely regulated via a complex genetic regulatory network incorporating both developmental and environmental factors. In an effort to identify negative regulators of leaf senescence, we screened activation-tagged Arabidopsis lines for mutants that exhibit a delayed leaf senescence phenotype. One of the mutants (ore7-1D) showed a highly significant delay of leaf senescence in the heterozygous state, leading to at least a twofold increase in leaf longevity. The activated gene (ORE7/ESC) encoded a protein with an AT-hook DNA-binding motif; such proteins are known to co-regulate transcription of genes through modification of chromatin architecture. We showed that ORE7/ESC, in addition to binding to a plant AT-rich DNA fragment, could also modify the chromatin architecture, as illustrated by an altered distribution of a histone–GFP fusion protein in the nucleus of the mutant. Globally altered gene expression, shown by microarray analysis, also indicated that activation of ORE7/ESC results in a younger condition in the mutant leaves. We propose that ectopically expressed ORE7/ESC is negatively regulating leaf senescence and suggest that the resulting chromatin alteration may have a role in controlling leaf longevity. Interestingly, activation of ORE7/ESC also led to a highly extended post-harvest storage life.
Leaf senescence is a developmentally programmed process that involves largely catabolic activities leading to the degeneration of the leaf. It is a highly regulated process that requires sequential and orchestrated changes in cellular physiology, biochemistry, and gene expression (Buchanan-Wollaston et al., 2003; Lim and Nam, 2005; Quirino et al., 2000). One of the distinctive changes that occurs during senescence is a rapid decline in photosynthetic activity. Other metabolic changes include hydrolysis of macromolecules that were accumulated during the growth phase followed by remobilization to the growing parts of plants, such as young leaves and reproductive organs. Therefore, leaf senescence is a critical phase in the plant life cycle that contributes to the fitness of plants by relocating and recycling nutrients (Bleecker and Patterson, 1997; Nam, 1997; Noodén, 1988).
Leaf senescence can be detrimental to yield and quality in agricultural crops and also contributes to post-harvest losses in vegetable and forage crops (Buchanan-Wollaston et al., 2003). Thus, increased understanding of the mechanisms controlling plant senescence should have great potential for agricultural improvement. Despite this, a genetic strategy to manipulate leaf senescence has been limited, in contrast to the case of fruit ripening where the benefits of controlling post-harvest characters have been well demonstrated. In order to control the timing and progression of senescence, the key regulatory molecules and pathways must be identified. Therefore, we have carried out a series of mutant screens to pinpoint genes that control leaf senescence in the model plant Arabidopsis (Kim et al., 2006; Oh et al., 1997; Woo et al., 2001, 2002).
Our initial screen for delayed leaf senescence mutants was focused on loss-of-function mutant pools in order to identify a positive regulator of senescence. However, redundancy between pathways means that loss of a regulatory element in one pathway might not significantly alter the senescence program, thus limiting the scope of this type of screening. We also expected that there might be cases in which senescence is controlled by negative regulators. In order to identify negative regulators of leaf senescence and to overcome the limitations associated with the loss-of-function mutant pool, we screened for delayed leaf senescence mutants in newly generated activation tagged pools of Arabidopsis. An advantage of this approach is that an activated gene usually behaves as a dominant, gain-of-function mutation, displaying the mutant phenotypes in the T1 generation.
In this paper we describe one delayed senescence mutant, ore7-1D, that was isolated from an activation tagging pool. This mutant exhibits markedly extended leaf longevity, due to increased expression of the ORE7 gene that encodes an AT-hook motif protein. In animals and yeast, many AT-hook motif proteins have been shown to have key roles in growth and development and act by modifying the chromosomal architecture to co-regulate transcription of a set of genes. We found that the ORE7 gene is identical to ESCAROLA (ESC) (Weigel et al., 2000), a gene that has been previously described but not functionally analyzed in planta. Through the analysis of the ore7-1D mutant, we suggest that regulation of chromatin structure might be important for the proper progression of leaf senescence. We also demonstrate that regulated expression of ORE7/ESC could be used to manipulate the post-harvest storage life of plants.
Isolation of the ore7 mutant
Approximately 20,000 activation tagged lines of Arabidopsis (Col-0) were generated in our laboratory using the pSKI105 vector (Weigel et al., 2000). Upon insertion into the plant genome, this vector allows activation of genes in its vicinity due to the presence of a strong enhancer element (4× 35S CaMV) in the vector. The generated pool was screened for delayed leaf senescence mutants by visually evaluating the degree of leaf yellowing at the T1 generation. This strategy enabled the identification of negative genetic regulators of leaf senescence, which exhibited delayed leaf senescence when overexpressed. Among the screened activation lines, the line named oresara7-1D (ore7-1D) exhibited the most noticeable delay of leaf senescence (Figure 1a). Segregation in the T2 generation indicated that the ore7 phenotype is due to a semi-dominant mutation caused by activation of a gene in the activation line (data not shown). In homozygous plants the whole developmental process, including leaf emergence and growth, was delayed, this causing difficulties in assessing the effect of the mutation on senescence in individual leaves (Hensel et al., 1993; Woo et al., 2001). However, in the heterozygous plants, where the expression of the activated gene is lower than that in the homozygous plants, the emergence time and growth rate of leaves was almost identical to those of the wild type. Therefore, the experiments described below were carried out using the heterozygous mutant. In addition to the delayed leaf senescence phenotype, the heterozygous ORE7/ore7-1D mutant showed slightly shorter petioles, round and enlarged leaves, an increased number of inflorescences, a late flowering phenotype and increased biomass (Figure 1a).
Extended leaf longevity in the ORE7/ore7-1D mutant
Visual examination of a single leaf throughout its life span revealed that in the ORE7/ore7-1D mutant leaf longevity was extended over twofold, which is one of the most significant extensions of leaf life span that has been reported in Arabidopsis (Figure 1b). Similarly the leaf chlorophyll content and the photochemical efficiency of photosystem II (PSII) (Fv/Fm) (Figure 1c,d), which are two physiological senescence markers, were both maintained at higher levels in the ORE7/ore7-1D mutant compared with the wild type. Moreover, in the mutant, a photosynthesis-related chlorophyll a/b-binding protein gene (CAB) was expressed at higher levels at later stages, while the induction of two senescence-associated genes, SEN4 and SAG12, was delayed (Figure 1e). The results together indicated that the ORE7/ore7-1D mutation results in a considerable delay of several parameters of leaf senescence and an extension of leaf longevity.
A gene encoding an AT-hook DNA-binding protein is overexpressed in the mutant
To identify the gene activated in the mutant, the genomic DNA fragment flanking the right border of the T-DNA insertion was isolated by plasmid rescue, utilizing the single EcoRI site in the T-DNA. The sequence of the genomic DNA flanking the T-DNA was then compared with genomic sequences of Arabidopsis, revealing a single predicted open reading frame of 936 bp closest to the right border. The predicted open reading frame (At1g20900) encodes a protein of 312 amino acids with a similarity to AT-hook DNA-binding proteins (Figure 2a) and was designated as the ORE7 gene. The ORE7 gene is identical to ESCAROLA (ESC) (Weigel et al., 2000), which was previously identified from an activation-tagged line which exhibited late flowering and altered leaf morphology. However, the in planta functions of this gene have not been previously analyzed.
Ribonucleic acid gel blot analysis revealed that expression of this gene was indeed activated in the mutant (Figure 2b). In addition, as expected from the difference in the phenotypic severity of the homozygous and heterozygous lines, the expression level of the activated gene was higher in the homozygous lines than in the heterozygous lines. Expression of the ORE7/ESC gene was low in wild-type plants, where it was not detectable by conventional RNA hybridization analysis. However, RT-PCR analysis indicated that the gene is down-regulated as leaf age increases (Figure 2c). This result is consistent with publicly available microarray data (Nottingham Arabidopsis Stock Centre).
To prove that the mutant phenotype observed in ORE7/ore7-1D was caused by overexpression of the predicted gene, two independent experiments were undertaken. Firstly, the delayed leaf senescence phenotype was re-created (ORE7/ore7-2D) by introducing the genomic DNA fragment from the activation line, containing the enhancer elements, a part of the 5′ upstream sequence, the whole open reading frame of the predicted gene, and a part of the 3′ flanking sequences, into wild-type plants (Figure1c,d). Secondly, a delayed leaf senescence phenotype was observed in independent transgenic lines where the full-length ORE7/ESC gene was expressed under the control of the 35S CaMV promoter (ORE7-OX1 and ORE7-OX2), although in this case the phenotype was less pronounced (Figure1c,d).
ORE7/ESC is a nuclear localized AT-hook protein
The predicted ORE7/ESC protein contains a nine amino acid domain, RPRGRPPGS, which satisfies the consensus requirement of an AT-hook motif (Figure 2a). Many of the AT-hook proteins characterized to date have been shown to either repress or activate transcription of a large number of genes by binding to AT-rich DNA sequences and modifying the architecture of DNA (Aravind and Landsman,1998; Sgarra et al., 2006; Strick and Laemmli,1995; Thanos and Maniatis, 1992). They are, therefore, referred to as architectural transcription regulators. In addition to the AT-hook motif, ORE7/ESC possesses stretches of glutamic acid, glycine, histidine, and glutamine, which are often found in transcriptional regulators (Manavathi et al., 2005; Zheng and Yang, 2004). If ORE7/ESC is an AT-hook DNA-binding protein, then it would be predicted to be localized in the nucleus. To examine the subcellular localization of ORE7, a green fluorescent protein (GFP)–ORE7 fusion protein was expressed in Arabidopsis protoplasts. As expected, ORE7–GFP was selectively localized to the nucleus (Figure 3a). Interestingly, the fusion protein appeared as distinctive nuclear speckles.
We then conducted an electrophoretic mobility shift assay (EMSA) to test whether the ORE7/ESC protein possessed the ability to bind to an A/T-rich DNA sequence, as would be expected from the proposed structure. The DNA sequence we used for this test was the A/T-rich region in the promoter of the pea (Pisum sativum) PRA2 gene (Nagano et al., 2001). As shown in Figure 3b (lane 2), ORE7 was able to bind to the 32P-labeled, 39-bp A/T-rich DNA fragment. Competitor DNA fragment (unlabeled) with the same wild-type sequence diminished the binding of ORE7 to the labeled fragment (lanes 3–5) more effectively than did the competitor DNA fragment with mutated sequences (lanes 6–8). Thus, binding of ORE7/ESC to this DNA fragment appeared specific. These results suggest that the ORE7/ESC gene encodes a protein that binds to AT-rich DNA sequences in the nucleus.
Chromatin architecture is modified in the ORE7/ore7-1D mutant
Use of a histone2B (H2B)–GFP fusion protein has been shown previously to be a sensitive method for the analysis of chromosome dynamics in living mammalian cells (Kanda et al., 1998). Histone2B–yellow fluorescent protein (YFP) was incorporated into the plant chromosome (Boisnard-Lorig et al., 2001) and shown to be useful for examining the status of plant chromatin. To monitor whether chromatin architecture is modified in the ore7-1D heterozygous mutants, we generated transgenic lines that express the H2B–GFP fusion protein and crossed them with the ORE7/ore7-1D mutants. The spatial distribution patterns of H2B–GFP-tagged chromatin of leaf cells are classified into three groups according to the number of subnuclear domains (nuclear bodies) and the interchromatin-like compartment that contains little or no GFP signal (Haithcock et al., 2005). While the leaf cells of ORE7/ore7-1D mutants exhibited more reticular chromatin distribution and intensely labeled nuclear bodies (Figure 3c, Group III), the majority of wild-type leaf cells displayed an evenly distributed fluorescence pattern over the whole nucleus with few or no nuclear bodies (Figure 3c, Group I). Therefore, increased levels of ORE7/ESC affect the distribution of H2B–GFP in the Arabidopsis nucleus, indicating that ORE7/ESC controls chromatin architecture.
Microarray analysis of gene expression in the ORE7/ore7-1D mutant
To gain an insight into the function of ORE7/ESC, a genome-wide gene expression analysis was performed by a microarray experiment. The Arabidopsis CATMA (Complete Arabidopsis Transcriptome Microarray) Consortium microarray (http://www.catma.org) (Allemeersch et al., 2005) containing a collection of gene specific tags (GSTs) of over 24000 Arabidopsis genes was probed with RNAs isolated from mature green leaves of the wild type or the ORE7/ore7-1D mutant. At this stage, neither the mutant nor the wild-type leaves show visible senescence symptoms. Sampling the leaves at this stage was to reduce inclusion of expression changes due to the secondary effect of the mutation in leaf senescence; later stage leaves have a larger difference in senescence state between the mutant and the wild type. The analysis revealed 1096 genes that showed at least a twofold change in expression in the mutant compared with the wild type. Among the 1096 genes, 615 genes (Table S1) and 481 genes (Table S2) were down- and up-regulated, respectively. A large number of genes that are up- or down-regulated during leaf senescence have been previously identified by using Affymetrix GeneChips (Buchanan-Wollaston et al., 2005). The information from these previous experiments was used to characterize the genes with altered expression in the ORE7/ore7-1D mutant. Since increased expression of ORE7 can delay leaf senescence, it was expected that senescence up-regulated genes (SEN+) would show lower expression in the ORE7/ore7-1D mutant. Conversely, senescence down-regulated genes (MG+) would show a higher expression in the mutant (the abbreviation MG+ is used here to denote genes that show higher expression at the mature green stage of a leaf). Among the 481 genes with a higher expression level in the ORE7/ore7-1D mutant, 92 (19%) and 39 (8%) were MG+ and SEN+, respectively. Among the 615 genes with a lower expression level in the ORE7/ore7-1D mutant, 195 (32%) and 42 (7%) were SEN+ and MG+, respectively (Tables S1 and S2). Thus, the gene expression pattern in the ORE7/ore7-1D mutant largely reflected a delayed leaf senescence phenotype. However, it is notable that, contrary to expectation, some SEN+ and MG+ genes show higher and lower expression, respectively, in the ORE7/ore7-1D mutant leaves.
Functional categories of genes with altered expression in the ORE7/ore7-1D mutant
Entering the array data into the MAPMAN program (Thimm et al., 2004) displayed the relative changes in expression of the various genes with an altered expression in the ORE7/ore7-1D mutant onto diagrams of the functional compartments (Figure 4a). Of particular interest was a large group of genes involved in protein synthesis that showed increased expression in ORE7/ore7-1D. For example, 16% (75 genes) of the up-regulated genes encoded ribosomal proteins (Figure 4, Table S1). In addition, various genes in the DNA synthesis group, including histone genes, were up-regulated. The increased expression of two histone genes and three ribosomal proteins was confirmed by RT-PCR analysis (Figure 4b).
In contrast, gene expression patterns indicate a reduction in activity of signaling pathways involving jasmonic acid (JA), abscisic acid (ABA), ethylene, and salicylic acid (SA) in the mutant (Figure 4, Table S2). Reduced expression of genes which depend on the SA signaling pathway for their expression, including those encoding a leucine-rich repeat family protein (At4g28490) and a protein kinase (At4g23150) (Buchanan-Wollaston et al., 2005), was observed. Also, the array results indicated that four JA pathway-related genes were down-regulated in the mutant (Table S2) and this was supported by RT-PCR analysis showing down-regulation of a plant defensin gene (PDF1.1, At1g75830), which is known to be dependent on JA signaling for its expression (Figure 4b). In addition, seven ethylene pathway genes were down-regulated in the mutant (Table S2) and this was confirmed by RT-PCR assay of an ethylene response factor (ERF1, At4g17500) (Figure 4b). Furthermore, RT-PCR showed that several ABA-inducible genes, including ABI1, rd29A, and rd29B were down-regulated in the mutant.
The differential expression of these pathways is not likely to be due to a large difference in the developmental stages between the mutant and wild-type plants, since we used mature green leaves from both plants and these showed no apparent senescence symptoms at that time. Also, there was no noticeable difference in the expression of photosynthesis-related genes, which would be expected if senescence had started in the wild-type leaves (Figure 4b, Tables S1 and S2).
Interestingly, while cytokinin is an effective senescence-retarding growth regulator (Gan and Amasino, 1995; Kim et al., 2006), expression of genes involved in cytokinin synthesis or signaling was not altered in the mutant (Figure 4b, Tables S1 and S2), indicating that the delayed senescence in the activation mutant line is not due to modification of cytokinin synthesis or signaling.
Ectopic ORE7/ESC delays leaf senescence induced by senescence-accelerating hormones
The gene expression data shown above revealed that the signaling pathways of ethylene, JA, and ABA that have a role in accelerating senescence were at least partially suppressed by ORE7/ESC expression in mature leaves. We thus examined the response of the mutant to these senescence-accelerating hormones (Figure 5). When detached leaves were treated with methyl jasmonate (MJ) or ABA a rapid decrease in chlorophyll and PSII activity was observed in wild-type leaves, and 3 days after incubation the chlorophyll content of wild-type leaves was less than 40%. In contrast, the ORE7/ore7-1D leaves retained more than 90% of chlorophyll following treatment with MJ or ABA. Following ethylene treatment, the difference in senescence response between the wild type and the mutant was not as striking as in the case of MJ or ABA but the mutant did retain more chlorophyll and photosynthetic activity than the wild type after 4 days of the treatment (Figure 5a). The altered senescence responses of the mutant to the hormones were further analyzed by measuring the expression of a senescence-induced gene, SEN4. This gene showed delayed expression in the mutant compared with the wild type following MJ and ABA treatment (Figure 5b), but little difference in response to ethylene was seen. These data showed that the senescence response pathways induced by MJ, ABA, and ethylene can be suppressed by ectopic ORE7/ESC, although the ethylene response was much less affected in our experimental conditions.
Increased post-harvest storage life of the ORE7/ore7-1D mutant
We then investigated the post-harvest storage life in leaves from the ORE7/ore7-1D mutant to see whether this mirrors the delay in the natural senescence of the leaf. Detached leaves were examined under conditions that accelerate senescence. Many stresses lead to premature senescence and are associated with generation of reactive oxygen species (Reilly et al., 2004). Therefore we assessed tolerance of the detached leaves to oxidative stress. The detached leaves were treated with hydrogen peroxide and a rapid decline of chlorophyll content and PSII activity was seen in the wild type. In contrast, the ORE7/ore7-1D leaves retained over 80% of chlorophyll (Figure 6a) and 95% of PSII activity (data not shown), even at 5 days after treatment. Darkness is a potent environmental stimulus that accelerates post-harvest senescence and, as shown in Figure 6(b), the post-harvest life of the detached leaves of the ORE7/ore7-1D mutant was significantly longer in this condition compared with wild type. Even after 8 days of dark incubation, over 50% of the chlorophyll content was retained in the ORE7/ore7-1D mutant, while wild-type leaves lost all chlorophyll by that time. In addition, the photochemical activity of PSII was maintained at a much higher level and induction of SEN4 expression, a marker gene for senescence, was lower in the mutant (data not shown). The data together showed that overexpression of ORE7, as in the ORE7/ore7-1D mutant, can be highly effective in delaying post-harvest senescence of detached leaves.
Constitutive expression of a gene in transgenic plants could cause several pleiotropic effects, as we observed in the ORE7/ore7-1D mutant. Therefore, the delayed senescence phenotype may be the result of an indirect downstream effect following the constitutive expression of the gene. To test this, we investigated whether controlled expression of the ORE7/ESC gene still leads to delayed senescence. For this experiment, we generated inducible transgenic lines that express the ORE7/ESC gene under the control of a glucocorticoid-inducible promoter. In the absence of dexamethasone (DEX, a synthetic glucocorticoid) treatment, no noticeable phenotypic alteration was observed in the transgenic plants, while expression of the ORE7/ESC gene was readily induced by treatment with DEX (Figure 6c). The effect of inducible overexpression of the ORE7/ESC gene on storage life of leaves was then assessed by the dark-induced post-harvest senescence assay using detached leaves. After 7 days of dark incubation, DEX-treated leaves of the transgenic ORE7/ESC inducible lines remained green, while control leaves turned yellow (Figure 6c). Measurement of the photochemical efficiency revealed that the DEX-treated leaves retained a higher photochemical efficiency than control leaves (Figure 6c). This result indicated that the delayed senescence in the detached leaves is a direct effect of ORE7/ESC on the senescence-controlling pathway. Importantly, exploiting a strategy of regulated expression of ORE7/ESC should be applicable to the manipulation of post-harvest senescence, thereby negating unfavorable pleiotropic effects.
The harvesting of vegetable crops often involves the whole of the aerial part of the plant. Therefore, we examined the post-harvest senescence response of the whole aerial part of the ORE7/ore7-1D activation mutant where the roots were removed from the mutant line during harvest as a simulation of vegetable harvest. Aerial parts of the plants harvested at 20 days after planting (DAP) by cutting the junction between the aerial part and the roots were kept on a wet filter for 6 days under long-day conditions. While the leaves of wild type became almost completely yellow by this time, the leaves of mutants remained green and fresh (Figure 6d). This result demonstrates that regulated ORE7/ESC overexpression could be a useful method for delaying post-harvest senescence at the whole plant level as well as the detached leaf level.
The timing of the onset and the rate of progression of leaf senescence must be finely controlled with concomitant regulation of gene expression during the senescence process. Studies utilizing molecular, genetic, and genomics strategies have enabled limited predictions of the molecular mechanisms that are involved in controlling leaf senescence (Buchanan-Wollaston et al., 2005; van der Graaff et al., 2006; Guo and Gan, 2006; Kim et al., 2006; Miao et al., 2004; Woo et al., 2001, 2002; Xiong et al., 2005). In this paper we present data showing that ectopic overexpression of ORE7/ESC, a member of the AT-hook protein gene family, greatly extends leaf longevity as well as post-harvest storage life in transgenic plants and suggest a possibility that this may occur by alteration of chromatin architecture.
This exciting discovery was enabled by the isolation of a new mutant from an activation-tagged mutant pool, which showed significantly delayed leaf senescence in the heterozygote state, ORE7/ore7-1D. Furthermore, leaf longevity was extended over twofold, which represents one of the most significant extensions of leaf life span so far reported in Arabidopsis. The altered senescence process of the ORE7/ore7-1D mutant is distinctly different from other previously identified leaf senescence mutants such as those caused by the ore1, ore3, ore4, ore9, or ore12 mutations (Oh et al., 1997; Woo et al., 2001, 2002). While these mutations bring about a delay mostly in the onset of timing of leaf senescence with relatively little effect on the progression of the senescence process, the ORE7/ore7-1D mutant showed a considerably slower progression of leaf senescence. This is similar to another mutation, dls1, which is also altered in the progression of leaf senescence (Yoshida et al., 2002). Interestingly, the ectopic expression of ORE7/ESC has a broad effect on senescence, causing a delayed senescence phenotype in organs other than the leaf, including the stem and silique (Figure S1). Furthermore, senescence induced by MJ, ABA, or ethylene is delayed as well as age-dependent leaf senescence.
Our initial screen for delayed leaf senescence mutants from activation tagging lines was attempted to identify a negative regulator of senescence. In order to confirm that ORE7 is a negative regulator of leaf senescence, we utilized the loss-of-function mutant. Since a T-DNA knockout line of ORE7/ESC was not available in the SALK collection, a TILLING strategy (for Targeting-Induced Local-scale Lesions In Genomes) was used to isolate mutants of the ORE7/ESC gene. We then examined the phenotype of the loss-of-function ore7/esc mutant, in which an early termination (converts Gln43 to stop codon; CS90175) occurs. The mutant did not exhibit any senescence phenotype, and no visible morphological phenotype was observed in the mutant either. This observation might reflect that the senescence phenotype we observed in the activation line could be a pleiotropic phenotype due to ectopic expression of the gene, not representing its primary in planta function. Analysis of the expression patterns of the ORE7/ESC gene indicates that it is expressed at a low level in leaves. Therefore, it may be that the role of this gene in normal plant development is not necessarily related to leaf senescence but the ectopic expression of the gene in leaves has a serendipitous effect resulting in delayed senescence. Alternatively, lack of phenotype in the ore7/esc mutant may be due to the presence of functionally redundant genes, since approximately 30 paralogs were identified in Arabidopsis.
The AT-hook motif is known to interact with the narrow minor groove of AT-rich DNA sequences. AT-hook proteins can affect the architecture of chromatin and play a role in the regulation of transcription by acting as factors that influence the association of transcription factors with chromatin (Aravind and Landsman, 1998; Sgarra et al., 2006; Strick and Laemmli, 1995; Thanos and Maniatis, 1992), although knowledge about the functional role of AT-hook motif proteins is still very limited in plants. From the data we presented here, we argue that ORE7 indeed functions as a plant AT-hook protein. It is localized in the nucleus and can bind to AT-rich sequences in vitro. ORE7/ESC can affect chromatin architecture, as we showed by modification of the distribution of H2B–GFP fusion (Figure 3). Furthermore, it can affect transcription of a large number of genes, as we demonstrated by microarray analysis. Studies on AT-hook proteins in wheat (Triticum aestivum) (AHM1, Morisawa et al., 2000) and Arabidopsis (AHL1, Fujimoto et al., 2004) have revealed that they can bind to the matrix attachment region (MAR) and are localized to the nuclear matrix, suggesting that they can function as a nuclear matrix component. ORE7/ESC appears to be paralogous to AHL1 with a similarity in sequence and protein size. However, ORE7/ESC has a feature different from AHL1. The Arabidopsis AT-hook motif proteins can be broadly classified into two phylogenic groups (Fujimoto et al., 2004). ORE7 and AHL1 belong to different groups. In addition, ORE7/ESC contains stretches of histidine, glutamic acids, and glutamine, which is a feature different from AHL1. Thus, ORE7/ESC and AHL1 may have distinctive in planta functions.
How then is leaf senescence delayed in the ORE7/ore7-1D mutant? With the notion that ORE7/ESC is an AT-hook protein, it is conceivable that ORE7/ESC can up-regulate genes that suppress senescence and down-regulate genes that enable the progression of the senescence process through modification of chromatin architecture. Indeed, the microarray data revealed enhanced expression of genes for self-maintenance as well as reduced expression of genes for senescence-promoting pathways in the activation line, although the genes with altered expression may not be the direct targets of ORE7/ESC. The possibility that altered chromatin architecture may control leaf longevity in Arabidopsis is analogous to the case of the yeast (Saccharomyces cerevisiae) SIR (Silent Information Regulator) gene. When this gene is overexpressed, the chromatin architecture of the yeast genome is altered, resulting in extended longevity (Campisi, 2000; Lin et al., 2000). However, the mechanisms suggested for the extended longevity are different. In yeast, chromatin compactness caused by overexpression of SIR was suggested to suppress age-dependent detrimental events, such as DNA rearrangement and rDNA circle formation. In the case of ORE7/ESC, we suggest that the extended longevity is caused by altered expression of a large number of genes that are up- or down-regulated during development.
Mutations that affect chromatin architecture often display complex and pleiotropic phenotypes, since in many cases they affect expression of a relatively large number of genes (Berger and Gaudin, 2003; Li et al., 2002). A similar situation was observed in the ORE7/ESC overexpressor lines (Figure 1). The pleiotropic phenotypes caused by overexpression of ORE7/ESC were consistent with the microarray analysis of the mutant. Ectopic ORE7/ESC affected not only expression of the senescence-related genes but also that of genes in many other functional categories. Unexpected up-regulation of some of the senescence-related genes also appears to be one of the pleiotropic effects of the ORE7/ESC overexpressor. It should be also noted that the delayed leaf senescence phenotype is not due to the indirect effect of delayed reproduction, i.e. late flowering, since, in Arabidopsis, senescence of individual leaves is not closely linked with the development of the reproductive structures (Hensel et al., 1993). This is also shown clearly by inducible expression experiments on the detached leaves.
Identification of genes that alter senescence has considerable practical value. In this paper, we showed that ORE7/ESC, when overexpressed in transgenic plants, has a profound effect in delaying post-harvest as well as natural senescence. The regulated expression of the ORE7/ESC gene could potentially be used for improving the storage life of harvested vegetables.
There is increasing evidence that chromatin structure plays a critical role in many facets of plant development (Berger and Gaudin, 2003; Li et al., 2002). Our results provide an important molecular insight that leaf longevity may be affected by the control of chromatin structure. The activation line will provide a valuable tool for defining the components and regulation of the pathways involved in this mechanism.
Plant materials, mutant screening, and growth conditions
Arabidopsis thaliana ecotype Columbia (Col-0) plants were grown in a temperature-controlled greenhouse at 24°C in 16-h days. Arabidopsis plants were transformed using Agrobacterium tumefaciens ABI that carries the activation-tagging plasmid pSKI015 by the floral dipping method (Clough and Bent, 1998). Mutants with a delayed leaf senescence phenotype were initially screened by visual evaluations of leaf yellowing of T1 plants. Potential mutants were self-fertilized and the progenies were further analyzed for co-segregation of the delayed senescence phenotype with the T-DNA insertion.
Assay of senescence
Age-dependent leaf senescence was assayed as described by Woo et al. (2001). For dark-induced post-harvest leaf senescence assay, the third and fourth leaves at 12 days after leaf emergence were detached and floated on 3 mm 2-(N-morpholine)-ethanesulfonic acid (MES) buffer (pH 5.7) in the dark. For oxidant treatment, detached leaves were floated in the MES buffer in the presence or absence of 15 mm hydrogen peroxide. For hormone treatments, detached leaves were floated in the MES buffer in the presence or absence of 50 μm ABA (Sigma, http://www.sigmaaldrich.com/) or 100 μm MJ (Sigma). In the case of ethylene treatment, detached leaves were incubated in a glass box containing 5 μm ethylene gas. All hormonal treatments were performed at 22°C under continuous light. The chlorophyll content of individual leaves was measured as described in Lichtenthaler (1987). The photochemical efficiency of PSII was measured by the Plant Efficiency Analyzer (Hansatech Instruments, http://www.hansatech-instruments.com/) (Oh et al., 1997). For assaying post-harvest senescence at the whole plant level, aerial parts of plants harvested at 20 DAP were cut out and placed on wet filter papers for 6 days.
Subcellular localization of the ORE7-GFP fusion protein
The full-length ORE7 open reading frame was amplified by PCR with primers containing appropriate restriction sites and cloned into the C-terminus of the GFP coding region in the p326GFP-3G vector. The plasmid DNA was introduced into protoplasts of wild-type leaf cells as described previously (Kim et al., 2006). Expression of fusion protein was observed 18 h after transfection using the Zeiss LSM 510 Meta confocal microscopy (Carl Zeiss, http://www.zeiss.com/).
Assay of chromatin architecture
The H2B–GFP fusion construct was generated to examine the effect of ORE7/ESC in the regulation of chromatin architecture. Transgenic plants expressing H2B–GFP under the control of cassava vein mosaic virus (CsVMV) promoter were generated by the floral dip method. These plants were crossed with the ORE7/ore7-1D heterozygote mutant. In the F2 generation, the plants that expressed H2B–GFP and showed the delayed senescence phenotype were selected by phenotype and fluorescence microscopy followed by PCR-based genotyping.
Changes in chromatin architectures were assessed by monitoring H2B–GFP fluorescence patterns (Boisnard-Lorig et al., 2001) in the first and second leaves of wild-type and the ORE7/ore7-1D mutant, respectively, using the LSM 5 live imaging confocal microscope (Carl Zeiss). The distinct H2B–GFP fluorescence patterns were scored and further classified into three groups to represent the changes in chromatin architectures. The data presented are the mean ± SE of each nucleus fraction from four independent experiments.
Electrophoretic mobility shift assay
The protein-coding sequence of the ORE7/ESC gene was amplified by PCR to introduce the restriction enzyme sites. The fragment was ligated into the corresponding restriction enzymes sites in the TriEx-1 vector (Novagen, http://splash.emdbiosciences.com/). In vitro protein synthesis was performed with 35S-labeled methionine (Perkin Elmer Life Science, http://las.perkinelmer.com/) in the TnT reticulocyte lysate system (Quick Coupled Transcription/Translation Systems; Promega, http://www.promega.com/) according to the manufacturer’s instructions.
An electrophoretic mobility shift assay was performed using a method described previously (Nagano et al., 2001). The 39-bp synthetic oligonucleotide (5′-TAA CAC ATA TTTTGATAA ATTTATTACTAA AACTATTTT-3′) that is derived from the pea PRA2 gene, a dark-induced small G protein gene (Yoshida et al., 1993), was used as a probe. The probe was labeled by incubation with [γ-32P]ATP and T4 polynucleotide kinase. Five microliters of TNT (in vitro transcription/translation)-produced ORE7 protein was added into a binding buffer (20 μl) containing 2 μg of poly(dI-dC)–poly(dI-dC), bovine serum albumin (500 μg μl−1). The protein–DNA complex was formed by incubating this mixture at 25°C for 1 h with the 32P-labeled probe. Competition experiments using the wild-type competitor and mutated competitors (5′-TAACACACTGCAGGATAA ATTTATTAC TAA AACTATTTT-3′) were also performed. Binding reactions were carried out with 5 μl of TNT-produced ORE7 protein and the 32P-labeled probe (2.5 fmol) in the presence of unlabeled wild-type competitor or mutant competitor.
Analysis of glucocorticoid-inducible lines
Transgenic lines expressing ORE7/ESC under the control of glucocorticoid-inducible promoter were generated by the floral dipping method. The glucocorticoid treatments were performed as described by Aoyama and Chua (1997). Detached mature green rosette leaves (12 days old) were floated in a buffer containing 15 μm DEX (Sigma) in the dark at 22°C for 7 days.
Microarrays used carried genome sequence tags (GST) fragments generated using gene-specific primers identified by the CATMA Consortium (http://www.catma.org) (Allemeersch et al., 2005). The CATMA PCR products were generated by secondary amplification as described on the CATMA web site. The PCR fragments were purified and mixed with DMSO to give a final DNA concentration of approximately 0.2 μg μl−1 in 50% DMSO. Microarrays were printed on CMT-GAPS coated slides (Corning, http://www.corning.com/) using a BioRobotics Microgrid II robot. Slides were baked for 4 h at 80°C and then stored with desiccant at room temperature.
For microarray hybridization, total RNAs were isolated from leaves of wild-type1 and ORE7/ore7-1D mutant at the age of 12 days (at mature green stage) and 1 μg of each RNA was amplified using the MessageAmp II aRNA Amplification kit (Ambion, http://www.ambion.com/) in accordance with the kit protocol with a single round of amplification. The Cy3- and Cy5-labelled cDNA probes were prepared from 2 μg of aRNA using the CyScribe Post-Labelling kit (Amersham Biosciences, http://www5.amershambiosciences.com/) whereby amino allyl-dUTP was incorporated during cDNA synthesis followed by chemical labeling of the amino allyl-modified cDNA using CyDye NHS-esters. Labeled cDNA was mixed in 25% formamide, 5× SSC, 0.1% SDS and 0.5 mg ml−1 yeast tRNA (Invitrogen, http://www.invitrogen.com/) and hybridized to slides overnight at 42°C. Slides were washed and then scanned using an Affymetrix 428 array scanner at 532nm (Cy3) and 635nm (Cy5). Scanned data were quantified using Imagene version 4.2 software (BioDiscovery, http://www.biodiscovery.com/). GeneSpring version 5.1 (Silicon Genetics, http://www.silicongenetics.com/) was used to normalize the data and identify genes showing significant differential expression in the mutant using the Benjamini and Hochberg multiple correction false discovery test. Ratio data were converted to a log base 2 scale and imported into MAPMAN (Thimm et al., 2004).
RT-PCR assays for expression of various genes
The expression profiles obtained from chip hybridizations were further validated by RT-PCR using first-strand cDNA synthesized from independently isolated RNA samples. First-strand cDNA was synthesized with 2 μg of total RNA by the ImProm-II Reverse Transcription system (Promega). All RT-PCR experiments were repeated at least twice.
We thank K. H. Suh, Y. S. Park, and B. H. Kim for excellent technical assistance. We thank Dr Detlef Weigel for providing the pSKI015 vector and Dr Roger Beachy for the pLAU6 plasmid. This work was supported by grants from the Crop Functional Genomics Research Program (to HGN; CG1312) and the MOST/KOSEF to the National Core Research Center for Systems Bio-Dynamics (R15-2004-033). The work by POL was supported by the Korea Research Foundation Grant funded by Korea Government (MOEHRD, Basic Research Promotion Fund, KRF-2005-261-C00075) and Technology Development Program for Agriculture and Forestry, Ministry of Agriculture and Forestry, Republic of Korea.